Electronic copy available at: http://ssrn.com/abstract=2641802

1

Data-driven, networked urbanism Rob Kitchin, NIRSA, Maynooth University, County Kildare, Ireland Rob.Kitchin@nuim.ie, @robkitchin

The Programmable City Working Paper 14

http://www.nuim.ie/progcity/

10th August, 2015

Prepared for Data and the City workshop, 31 Aug-1st Sept 2015, Maynooth University Abstract

For as long as data have been generated about cities various kinds of data-informed urbanism

have been occurring. In this paper, I argue that a new era is presently unfolding wherein

data-informed urbanism is increasingly being complemented and replaced by data-driven,

networked urbanism. Cities are becoming ever more instrumented and networked, their

systems interlinked and integrated, and vast troves of big urban data are being generated and

used to manage and control urban life in real-time. Data-driven, networked urbanism, I

contend, is the key mode of production for what have widely been termed smart cities. In

this paper I provide a critical overview of data-driven, networked urbanism and smart cities

focusing in particular on the relationship between data and the city (rather than network

infrastructure or computational or urban issues), and critically examine a number of urban

data issues including: the politics of urban data; data ownership, data control, data coverage

and access; data security and data integrity; data protection and privacy, dataveillance, and

data uses such as social sorting and anticipatory governance; and technical data issues such as

data quality, veracity of data models and data analytics, and data integration and

interoperability. I conclude that whilst data-driven, networked urbanism purports to produce

a commonsensical, pragmatic, neutral, apolitical, evidence-based form of responsive urban

governance, it is nonetheless selective, crafted, flawed, normative and politically-inflected.

Consequently, whilst data-driven, networked urbanism provides a set of solutions for urban

problems, it does so within limitations and in the service of particular interests.

Key words: big data, data analytics, governance, smart cities, urban data, urban informatics,

urban science

Electronic copy available at: http://ssrn.com/abstract=2641802

2

Introduction

There is a rich history of data being generated about cities concerning their form, their

citizens, the activities that take place, and their connections with other locales. These data

have been generated in a plethora of different ways, including audits, cartographic surveying,

interviews, questionnaires, observations, photography, and remote sensing, and are

quantitative and qualitative in nature, stored in ledgers, notebooks, albums, files, databases,

and other media. Data about cities provide a wealth of facts, figures, snapshots and opinions

that can be converted into various forms of derived data, transposed into visualisations, such

as graphs, maps, and infographics, analyzed statistically or discursively, and interpreted and

turned into information and knowledge. As such, urban data form a key input for

understanding city life, solving urban problems, formulating policy and plans, guiding

operational governance, modelling possible futures, and tackling a diverse set of other issues.

For as long as data have been generated about cities then, various kinds of data-informed

urbanism have been occurring.

A new era is, however, presently unfolding wherein data-informed urbanism is

increasingly being complemented and replaced by data-driven, networked urbanism. Here,

urban operational governance and city services are becoming highly responsive to a form of

networked urbanism in which big data systems are prefiguring and setting the urban agenda

and are influencing and controlling how city systems respond and perform. In short, we are

moving into an era where cities are becoming ever more instrumented and networked, their

systems interlinked and integrated, and the vast troves of data being generated used to

manage and control urban life. Computation is now routinely being embedded into the fabric

and infrastructure of cities that, on the one hand, is producing a deluge of contextual and

actionable data, and on the other acts on such data in real-time. Moreover, data that used to

be the preserve of a single domain are increasingly being shared across systems enabling a

more holistic and integrated view of city services and infrastructures. As such, cities are

becoming knowable and controllable in new dynamic ways, responsive to the data generated

about them. I thus argue that data-driven, networked urbanism is the key mode of production

for what have widely been termed smart cities.

In this paper I provide a critical overview of data-driven, networked urbanism

focusing in particular in particular on the relationship between data and the city, rather than

network infrastructure, computational or urban issues. The paper starts by setting out how

cities are being instrumented and captured as big urban data, how these data are being used to

manage and control cities, and how data-driven, networked urbanism is underpinning the

3

emergence of smart cities. This is then followed by a critical examination of a number of

problematic issues related to data-driven, networked urbanism, including: the corporatisation

of governance (data ownership, data control, data coverage and access); the creation of

buggy, brittle, hackable urban systems (data security, data integrity); social, political, ethical

effects (data protection and privacy, dataveillance, and data uses including social sorting and

anticipatory governance); and technical data issues (data quality; veracity of urban data

models and data analytics; data integration and interoperability).

Big data and smart cities

Since the start of computing era urban data have been increasingly digital in nature, either

digitized from analogue sources (manually entered or scanned) or born digital, generated by

digital devices, stored as digital files and databases, and processed and analyzed using

various software systems such as information management systems, spreadsheets and stats

packages, and geographic information systems. From the 1980s onwards, public

administration records, official statistics, and other forms of urban data were released

predominately in digital formats and processed and analyzed through digital media.

However, these data were (and continue to be) generated and published periodically and often

several months after generation.

In cases such as exhaustive datasets - for example, detailed maps or national censuses

- new surveys are very infrequent (e.g., 10 years for censuses) and their publication might be

18-24 months after collection, and longer for specific subsets. For domain specific issues,

such as transport and traffic flows or public transportation usage, surveys are conducted every

few years, using a limited spatial and temporal sampling framework. Only a handful of

datasets are published monthly (e.g. unemployment rates) or quarterly (e.g. GDP), with most

being updated annually due to the effort required to generate them. These data typically have

poor spatial resolution, referring to large regions or the nation, and little disaggregation (e.g.,

by population classes or economic sectors). In cases where data generation is more frequent,

such as remote sensing, only occasional snapshots are bought by city administrations due to

their licensing costs. In other cases, such as consumer purchasing (as evidenced in credit

card transactions) the data was largely black-boxed within a financial institution. In other

words, whilst there was a range of urban digital data available to urban managers and policy

makers from the 1980s through to 2000s , along with increasingly sophisticated software such

as GISs to make sense of them, sources of data were temporally, spatially and domain (scope)

limited.

4

Post-Millennium, the urban data landscape has been transformed, with a massive step

change in the nature and production of urban data, transitioning from small data to big data,

wherein the generation of data is continuous, exhaustive to a system, fine-grained, relational,

and flexible (see Table 1) across a range of domains (Kitchin 2014a). From a position of

relative data scarcity, the situation is turning to one of data deluge. This is particularly the

case with urban operational data wherein traditional city infrastructure, such as transportation

(e.g., roads, rail lines, bus routes, plus the vehicles/carriages) and utilities (e.g., energy, water,

lighting), have become digitally networked, with grids of embedded sensors, actuators,

scanners, transponders, cameras, meters and GPS producing a continuous flow of data about

infrastructure conditions and usage (constituting what has been called the Internet of Things).

Many of these systems are generating data at the individual level, tracking individual travel

passes, vehicle number plates, mobile phone identifiers, faces and gaits, buses/trains/taxis,

meter readings, etc (Dodge and Kitchin 2005). These are being complemented with big data

generated by: (a) commercial companies such as mobile phone operators (location, app use),

travel and accommodation sites (reviews), social media sites (opinions, photos, personal info,

location), transport providers (routes, traffic flow), website owners (clickstreams), financial

institutions and retail chains (purchases), and private surveillance and security firms

(location, behaviour) that are increasingly selling and leasing their data through data brokers,

or making their data available through APIs (such as Twitter and Foursquare); (b)

crowdsourcing (e.g., Open Street Map) and citizen science (e.g., personal weather stations)

initiatives, wherein people collaborate on producing a shared data resource or volunteer data.

Other kinds of more irregular urban big data include digital aerial photography via planes or

drones, or spatial video, LiDAR (light detection and ranging), thermal or other kinds of

electromagnetic scans of environments that enable the mobile and real-time 2D and 3D

mapping of landscapes. And whilst official statistics are largely still waiting to undergo the

data revolution (Kitchin 2015), the generation of public administration data has been

transformed through the use of e-government online transactions that produce digital data at

the point-of-collection.

Table 1: Comparing small and big data

Small data Big data Volume Limited to large Very large Velocity Slow, freeze-framed/bundled Fast, continuous Variety Limited to wide Wide Exhaustivity Samples Entire populations Resolution and identification Course & weak to tight & strong Tight & strong

5

Relationality Weak to strong Strong Flexible and scalable Low to middling High Source: Kitchin (2014b) We are at start of this new big data era and the flow and variety of urban data is only

going to grow and diversify. Moreover, whilst much of these data presently remain in silos

and are difficult to integrate and interlink due to varying standards and formats, they will

increasingly be corralled into centralised systems such as inter-agency control rooms for

monitoring the city as a whole (e.g., Centro De Operacoes Prefeitura Do Rio in Rio de

Janeiro, Brazil, a data-driven city operations centre that pulls together into a single location

real-time data streams from thirty agencies, including traffic and public transport, municipal

and utility services, emergency and security services, weather feeds, information generated

by employees and the public via social media, as well as administrative and statistical data,

and is overseen by a staff of 180 data operatives -- see Figure 1 for examples of urban control

rooms), or what have been termed City Operating Systems (or City OS, such Microsoft’s

CityNext, IBM’s Smarter City, Urbiotica’s City Operating System, and PlanIT’s Urban

Operating System; see Figure 2). The latter are effectively Enterprise Resource Planning

(ERP) systems designed to coordinate and operate the activities of large companies

repurposed for cities. With the advent of the open data movement some of these data will

also feed into public-facing urban dashboards that provide a mix of interactive visualisations

of real-time, public administration and official statistical data (Kitchin et al. 2015a, see Figure

3).

Further, the production of these new big data have been accompanied by a suite of

new data analytics designed to extract insight from very large, dynamic datasets, consisting

of four broad classes: data mining and pattern recognition; data visualization and visual

analytics; statistical analysis; and prediction, simulation, and optimization (Miller 2010;

Kitchin 2014b). These analytics rely on machine learning (artificial intelligence) techniques

and vastly increased computational power to process and analyze data. Moreover, they

enable a new form of data-driven science to be deployed that rather than being theory-led

seeks to generate hypotheses and insights ‘born from the data’ (Kelling et al. 2009). This is

leading to the development of ‘urban informatics’ (Foth 2009), an informational and human-

computer interaction approach to examining and communicating urban processes, and ‘urban

science’, a computational modelling approach to understanding and explaining city processes

that builds upon and radically extends quantitative forms of urban studies that have been

6

Figure 1: Urban control rooms (Rio de Janeiro, Sydney, Glasgow and London)1

Figure 2: City Operating Systems (Microsoft CityNext, IBM Smarter Cities, Urbiotica City Operating System and PlanIT Urban Operating System) 2

7

Figure 3: Urban dashboards (Dublin, London, Amsterdam)3

practised since the 1950s, blending in geocomputation, data science and social physics (Batty

2013). Whereas urban informatics is more human-centred, interested in understanding and

facilitating the interactions between people, space and technology, urban science promises to

not only make sense of cities as they presently are (by identifying relationships and urban

‘laws’), but to also predict and simulate likely future scenarios under different conditions,

potentially providing city managers with value insight for planning and development

decision-making and policy formulation.

Urban big data, city operating systems, urban informatics, and urban science analytics

provide the basis for a new logic of urban control and governance -- data-driven, networked

urbanism -- that enables real-time monitoring and steering of urban systems and the creation

of what has widely been termed smart cities. The notion of a smart city can be traced back to

experiments with urban cybernetics in the 1970s (Flood 2011; Townsend 2013), the

development of new forms of city managerialism and urban entrepreneurship, including

smart growth and new urbanism, in the 1980s and 90s (Hollands 2008, Wolfram 2012,

Söderström et al., 2014, Vanolo 2014), and the fusing of ICT and urban infrastructure and

development of initial forms of networked urbanism from the late 1980s onwards (Graham

8

and Marvin 2001, Kitchin and Dodge 2011). As presently understood, a smart city is one that

strategically uses networked infrastructure and associated big data and data analytics to

produce a:

• smart economy by fostering entrepreneurship, innovation, productivity,

competiveness, and producing new forms of economic development such as the app

economy, sharing economy and open data economy;

• smart government by enabling new forms of e-government, new modes of operational

governance, improved models and simulations to guide future development, evidence-

informed decision making, better service delivery, and making government more

transparent, participatory and accountable;

• smart mobility by creating intelligent transport systems and efficient, inter-operable

multi-modal public transport;

• smart environments by promoting sustainability and resilience and the development of

green energy;

• smart living by improving quality of life, increasing safety and security, and reducing

risk;

• smart people by creating a more informed citizenry and fostering creativity,

inclusivity, empowerment and participation (Cohen 2012; Hollands 2008; Townsend

2013).

In short, the smart city promises to solve a fundamental conundrum of cities – how to reduce

costs and create economic growth and resilience at the same time as producing sustainability

and improving services, participation and quality of life – and to do so in commonsensical,

pragmatic, neutral and supposedly apolitical ways by utilising a fast-flowing torrent of urban

data and data analytics, algorithmic governance, and responsive, networked urban

infrastructure. Moreover, much more information is being placed into the hands of the public

to aid decision-making, navigation and participation through a plethora of locative social

media (apps that tell them about the city and which they can contribute to), open data sites,

public dashboards, hackathons, and so on.

The notion of smart cities, and the mode of data-driven, networked urbanism, has not,

however, been universally welcomed and has been subject to a number of critiques.

9

First, smart city initiatives treat cities as a set of knowable and manageable systems that act in

largely rational, mechanical, linear and hierarchical ways and can be steered and controlled

(Kitchin et al., 2015). Second, smart city initiatives are largely ahistorical, aspatial and

homogenizing in their orientation and intent, treating cities as if they are all alike in terms of

their political economy, culture, and governance (Greenfield 2013). Third, an emphasis is

placed on creating technical rather political/social solutions to urban problems thus overly

promoting technocratic forms of governance (Moronez 2013). Fourth, the project of

producing smart cities tends to reinforce existing power geometries and social and spatial

inequalities rather than eroding or reconfiguring them (Datta 2015). Fifth, the approach fails

to recognize the politics of urban data and the ways in which they are the product of complex

socio-technical assemblages (Kitchin 2014b). Sixth, the smart city agenda is being overly

driven by corporate interests who are using it to capture government functions as new market

opportunities (Hollands 2008). Seventh, networking city infrastructure potentially creates

buggy, brittle, and hackable urban systems (Kitchin and Dodge 2011; Townsend 2013). And

finally, data-driven, networked urbanism produces a number of activities that have profound

social, political, ethical consequences, including dataveillance and extensive geosurveillance,

social and spatial sorting, and anticipatory governance (Graham 2005; Kitchin 2014a).

In the rest of this paper, I want to concentrate on the last four critiques, and in

particular their associated data issues (rather than other aspects of the technical stacks of

urban socio-technical assemblages, and wider political-economic framing and effects),

including technical data issues, as way of further illustrating some of the challenges posed by

data-driven, networked urbanism and the need to further examine the relationship between

data and the city.

Data and the City

The politics of urban data

One of the key arguments for adopting a data-driven approach to urban governance is that it

provides a strong evidence-based approach to decision-making, system control, and policy

formation, rather than one that is heavily anecdotal, clientelist or localist. How an urban

system/infrastructure is run is thus less open to political influence and instead is driven by

objective, neutral facts in a technocratic, commonsensical, pragmatic way. Technical

systems and the data they produce are objective and non-ideological and thus politically

benign. Sensors, networked infrastructure, and computers have no inherent politics -- they

simply measure a value, communicate those values, and process, analyze and display the data

10

using scientific principles; producing measurements, records and information that reflect the

truth about cities. And while data from social systems, such as social media platforms (e.g.,

Twitter), are inherently more subjective and noisy, they provide a direct reflection of the

views, interactions and behaviour of people, in contrast to official surveys which reflect what

people say they do or think (or what they think the surveyor wants to hear), thus providing

better ground truthing of social reality. As such, big data about cities can be taken at face

value and used unconditionally shed light on cities and to manage and control urban systems

and infrastructure and guide urban policy.

The reality is somewhat different for two reasons. First, there are a number of

technical issues concerning data coverage, access and quality that means that the view data

presents of the city is always partial and subject to caution. Second, data are the products of

complex socio-technical assemblages that are framed and shaped by a range of technical,

social, economic and political forces and are designed to produce particular outcomes

(Kitchin 2014b; see Figure 4). On the one hand, what data are produced, how they are

handled, processed, stored, analyzed and presented is the result of a particular technical

configuration and how it is deployed (e.g., where sensors are located, their field of view, their

sampling rate, their settings and calibration, etc). On the other, how a system is designed and

run is influenced by systems of thought, technical know-how, the regulatory environment,

funding and resourcing, organisational priorities and internal politics, institutional

collaborations, and marketplace demand. In other words, a data assemblage possesses a

‘dispositif’, defined by Foucault (1977: 194) as a: ‘thoroughly heterogeneous ensemble

consisting of discourses, institutions, architectural forms, regulatory decisions, laws,

administrative measures, scientific statements, philosophical, moral and philanthropic

propositions.’ For Foucault, a dispositif is inherently political producing what he terms

‘power/knowledge’, that is knowledge that fulfils a strategic function. In other words, urban

big data are never neutral and objective, but rather are situated, contingent, relational, and

framed and used contextually to try and achieve certain aims and goals (to monitor, enhance,

empower, discipline, regulate, control, produce profit, etc.). Or to put it another way, urban

data are never raw but are always already cooked to a particular recipe for a particular

purpose (Bowker 2005; Gitelman 2013). As such, data-driven, networked urbanism is

thoroughly political seeking to produce a certain kind of city. It is thus necessary when

examining urban big data to critically unpack their associated data assemblage (including the

entire technical stack - infrastructure, platform, software/algorithms, data, interface) to

11

document how it is constituted and works in practice to produce urban processes and

formations, and for whose benefit.

Figure 4: A data assemblage

Data access, data ownership and data control

As already noted, much of the data presently being generated about cities are produced by

commercial companies, such as mobile phone operators, and private utility and transport

companies. For them, their data are a valuable commodity that provides competitive

advantage or an additional revenue stream if sold/leased, and they are under no obligation to

share freely the data they generate through their operations with city managers or the public.

As noted in 2014 by the British Minister for Smart Cities, Dan Byles MP4, the privatisation of

public services in the UK and elsewhere has also meant the privatisation of their associated

data unless special provision was made to ensure it was shared with the city or made open.

Similarly, access to data within public-private partnerships and semi-state agencies, or state

agencies operating as trading funds (such as the Met Office and Ordnance Survey in the UK

who generate significant operating costs by selling data and services), can be restricted or

costly to purchase. Consequently key framework datasets (e.g., detailed maps) can have

limited access and data concerning transportation (e.g., bus, rail, bike share schemes, private

tolls), energy, and water be entirely blackboxed. Even within the public sector, data can be

siloed within particular departments and not be shared with other units within the

organisation, or be open for other institutions or the public to use. As such, whilst there

might be a data revolution underway, access to much of that data is limited, and there are a

12

number of issues that need to be explored with respect to data ownership and data control,

especially with respect to procurement and the outsourcing or privatisation of city services.

Moreover, even if all data were to be open and shared it needs to be acknowledged that there

are still many aspects about cities where data generation is weak or absent. For example, in a

recent audit of Dublin datasets to determine whether the city was in a position to apply for

ISO37120 (the ISO standard for city indicators) data could only be sourced for 11 of 100

indicators sought (predominately because the data sought was either privatised or released at

an inappropriate scale).

Data security and data integrity

One of the prime anxieties of networking infrastructure and ubiquitous urban computing is

the creation of systems and environments which are inherently buggy and brittle and are

prone to viruses, glitches, crashes, and security hacks (Kitchin and Dodge 2011; Townsend

2013). As Mims (2013) notes, any networked device is open to be hacked and its data stolen

and used for criminal purposes, or corrupted, or controlled remotely, or misdirected, or to spy

on its users. The media report almost daily on large-scale data breaches of commercial

companies and state agencies and the theft of valuable personal data, with several incidents of

city infrastructure such as traffic management systems being hacked, disabled and controlled

(Paganini 2013). As Townsend (2013) notes, the notion of smart cities takes two open,

highly complex and contingent systems ― cities and computing ― and binds and networks

them together, meaning that data-driven, networked urbanism has in-built vulnerabilities.

Moreover, as urban systems evolve to become more complex, interconnected and

interdependent these vulnerabilities potentially multiply (Townsend 2013). Creating secure

big urban data systems is thus set to be a significant on-going task if public trust in their

purported benefits are to be gained and maintained. Another significant element in upholding

trust in data-driven, networked urbanism is how and what purposes the data are deployed.

Data uses

Urban big data are presently being used to undertake a diverse range of tasks, some of which

seem relatively benign, such as monitoring city lighting with the aim of improving the quality

of light and reducing its cost, and others more clearly political, such as directing policing

activity. A significant concerns is that as more and more data about cities and their citizens

are generated, privacy becomes eroded. Privacy is considered a basic human right, a

condition that people expect and value in developed countries. Yet, as sensors, cameras,

13

smartphones, and other embedded and mobile devices generate evermore data it becomes

increasingly difficult to protect, with individuals leaving ever greater quantities of digital

footprints (data they themselves leave behind) and data shadows (information about them

generated by others). Such troves of data are amenable to dataveillance, a mode of

surveillance enacted through sorting and sifting datasets in order to identify, monitor, track,

regulate, predict and prescribe (Clarke 1988; Raley 2013), and geosurveillance, the tracking

of location and movement of people, vehicles, goods and services and the monitoring of

interactions across space (Crampton 2003). Given the always-on nature of many of these

systems, and the tracking of unique identifiers, such dataveillance and geosurveillance are

becoming continuous and fine-grained with, for example, mobile phone companies always

knowing the location of a phone whilst it is not powered down (Dodge and Kitchin 2005).

Moreover, as data minimization norms become relaxed there are anxieties that data are being

shared, combined and used for purposes for which they were never intended. In particular,

the last twenty years have witnessed the rapid growth of a number of data brokers who

capture, gather together and repackage data for rent (for one time use or use under licensing

conditions) or re-sale, and produce various derived data and data analytics.

Whilst focusing on different markets, these data brokers seek to mesh together off-

line, online and mobile data to provide comprehensive views of people and places and to

construct personal and geodemographic profiles (Goss 1995; Harris et al., 2005). These

profiles are then used to predict behaviour and the likely value or worth of an individual and

to socially sort them with respect to credit, employment, tenancy and so on (Graham 2005).

The concern is that these firms practice a form of ‘data determinism’ in which individuals are

not profiled and judged just on the basis of what they have done, but on the prediction of

what they might do in the future using algorithms that are far from perfect, and yet are black-

boxed and lack meaningful oversight and remediate procedures (Ramirez 2013). Such

anticipatory governance can have far reaching effects. For example, a number of US police

forces are now using predictive analytics to anticipate the location of future crimes and direct

patrols, and to identify individual most likely to commit a crime in the future, designating

them pre-criminals (Stroud 2014). In such cases, a person’s digital footprints and data

shadow does more than follow them; it precedes them. Data assemblages then do not act as

cameras reflecting the world as it is, but rather as engines shaping the world in diverse ways

(Mackenzie 2008).

Technical data issues

14

Beyond data always being political and often restricted in access and limited in scope, it is

important to recognize that there are also a number of technical issues that place constraints

on the extent to which cities are knowable and controllable. Generating data is always an

open process. Approaches, methodologies, procedures, standards, and equipment are

designed, tested, negotiated, and debated. The data produced are shaped by technical

instruments, protocols, scientific norms, and scientist behaviour and organisational processes,

meaning they contain instrument and human error and bias. Moreover, generating data

always involves a process of abstraction (capturing particular measurements from the sum of

all possible data), representation (converting what is being measured into a readable form

(e.g., numbers, a wave pattern, a scatterplot, a stream of binary code, etc.)), and often

generalisation (e.g., into a set of categories) or calibration (transformed to compensate for

suspected error/bias). With any dataset then there are questions concerning data veracity and

quality and how accurately (precision) and faithfully (fidelity) the data represent what they

are meant to (especially when using samples and proxies), and how clean (error and gap

free), untainted (bias free), consistent (few discrepancies), and reliable (the measurement

instrument consistently produces the same quality of results) the data are (Goodchild 2009,

Kitchin 2014b). Further, because data are generated in so many different ways, using a

plethora of instruments and standards, it remains difficult to join them together to produce a

more holistic picture. As such, it impossible to measure the ‘truth’ of cities, but rather only

generate partial, selected views from particular vantage points. And what those views show

can be dirty, gamed, and faked.

Likewise, urban data models are created, with ontologies constructed rather than

essential (existing as a natural truth) and data analytics are selected, different parameters

selected and tweaked, and protocols applied. There are therefore questions as to the veracity

of models and analytics and the extent to which they shape the findings produced. To be

clear, urban informatics and urban science are seeking to produce as much insight as possible,

with as much validity as achievable, and they do provide useful knowledge. However, they

nonetheless produce a particularised vision and explanation of the city. Moreover, the output

they produce is open to misinterpretation and ecological fallacies. With respect to cities, one

of the most common types of ecological fallacy is introduced by the Modifiable Areal Unit

Problem (MAUP) (Openshaw 1984), wherein the statistical geography used to display

aggregate data can have a marked effect on the pattern of observations, and thus the

conclusions drawn (see Figure 5). Likewise, altering the classification boundaries, or altering

the number of classes, can have a similar effect. How data are classified and the scale at

15

which they are displayed can thus have a dramatic effect on how a city is understood and how

this feeds into how it is governed. While these effects are well understood by academic

statisticians, they are much less so within the policy community, and they are mostly

overlooked or ignored in applied research.

Figure 5: Mapping the same data at three different administrative scales

Conclusion

We are entering an era where computation is being routinely embedded into urban

environments and networked together, and people are moving about with smartphones that

ensure always available connectivity and access to information. These devices and

infrastructures are producing and distributing vast quantities of data in real-time, and they are

also responsive to these data and the analytics undertaken on them enabling new kinds of

monitoring, regulation and control. Cities then are becoming data-driven and are enacting

new forms of algorithmic governance. However, the data and algorithms underpinning them

are far from objective and neutral, but rather are political, imperfect, and partial. The smart

cities that data-driven, networked urbanism purports to create are then smart in a qualified

sense. Their production and operation is based on much more data and derived information

than previous generations of urbanism, but it is a form of urbanism that is nonetheless still

selective, crafted, flawed, normative and politically-inflected. Moreover, while the

instrumental rationality of data-driven, networked urbanism promotes urban knowledge and

management rooted in a quite narrowly framed ‘episteme (scientific knowledge) and teche

(practical instrumental knowledge)’, it is important that other forms of knowing, such as

‘phronesis (knowledge derived from practice and deliberation) and metis (knowledge based

on experience)’ (Parsons 2004: 49) are not silenced, providing both a counter-weight to the

limits of smart cities and positions from which to reflect on, critique and recast the production

of data-driven, networked urbanism. Indeed, whilst data-driven, networked urbanism

undoubtedly provides a set of solutions for urban problems, we also have to recognize that it

has a number of shortcomings and a number of potential perils. The challenge facing urban

16

managers and citizens in the age of smart cities is realise the benefits of planning and

delivering city services using a surfeit of data, evidence and real-time responsive systems

whilst minimizing any pernicious effects. To do that we have to be as smart about data and

data analytics as we would like to be about cities.

Acknowledgements

The research for this paper was provided by a European Research Council Advanced

Investigator Award, ‘The Programmable City’ (ERC-2012-AdG-323636).

Notes

1. Sources of figures: top-left: http://ipprio.rio.rj.gov.br/centro-de-operacoes-rio-usa-mapas-feitos-pelo-ipp/; bottom-left: http://www.dailytelegraph.com.au/news/nsw/sydney-under-watch-new-cameras-in-the-wake-of-thomas-kellys-kinghit-death/story-fni0cx12-1226777921307; top- right: http://www.eveningtimes.co.uk/news/13273769.Officers_get_the_whole_picture_at_new_centre/; bottom-right: http://archinect.com/news/article/75835110/who-s-your-data-urban-design-in-the-new-soft-city

2. Sources of figures: top-left: http://www.urenio.org/2013/10/22/microsoft-citynext/; bottom-left: http://www.urenio.org/2011/06/29/ibm-redbooks-smarter-cities-series/; top-right: http://www.plataformaarquitectura.cl/cl/02-308620/nuevo-contexto-urbano-espacios-publicos-flexibles-10-principios-basicos; bottom-right: http://raptorsme.tumblr.com/post/42271513343/the-uos-tm-architecture-the-controls-and

3. Sources of figures: left: http://www.dublindashboard.ie/; top-right: http://citydashboard.org/london/; bottom-right: http://visual.ly/city-dashboard-amsterdam

4. https://www.youtube.com/watch?v=3E3RpGMKbhg

References

Batty, M. (2013) The New Science of Cities. MIT Press, Cambridge, MA.

Bowker, G. (2005) Memory Practices in the Sciences. MIT Press, Cambridge, MA.

Clarke, R. (1988) Information Technology and Dataveillance. Communications of the ACM:

31(5): 498-512.

Cohen, B. (2012) What Exactly Is A Smart City? Fast Co.Exist, Sept 19th 2012,

http://www.fastcoexist.com/1680538/what-exactly-is-a-smart-city (last accessed 28 April

2015)

Crampton, J. (2003) Cartographic Rationality and the Politics of Geosurveillance and

Security. Cartography and Geographic Information Science 30(2): 135-148.

Datta, A. (2015) New urban utopias of postcolonial India: ‘Entrepreneurial urbanization’ in

Dholera smart city, Gujarat, Dialogues in Human Geography, 5(1): 3-22.

17

Dodge, M. and Kitchin, R. (2005) Codes of life: Identification codes and the machine-

readable world. Environment and Planning D: Society and Space. 23(6): 851 – 881

Gitelman, L. (ed) (2013) “Raw Data” is an Oxymoron. MIT Press, Cambridge.

Goss, J. (1995) ‘We know who you are and we know where you live’: the instrumental

rationality of geodemographics systems. Economic Geography 71: 171-198.

Graham, S. (2005) Software-sorted geographies. Progress in Human Geography 29: 562-80.

Graham, S., and Marvin, S. (2001) Splintering Urbanism: Networked Infrastructures,

Technological Mobilities and the Urban Condition. Routledge, New York.

Greenfield, A. (2013) Against the Smart City. New York: Do Publications.

Flood, J. (2011) The Fires: How a Computer Formula, Big Ideas, and the Best of Intentions

Burned Down New York City--and Determined the Future of Cities. Riverhead, New York.

Foth, M. (ed) (2009) Handbook of Research on Urban Informatics: The Practice and

Promise of the Real-Time City. IGI Global, New York.

Harris, R. Sleight, P. and Webber, R. (2005) Geodemographics, GIS and neighbourhood

targeting. Wiley, Chichester.

Hollands, R.G. (2008): Will the real smart city please stand up? City 12:3, 303-320

Kelling, S., Hochachka, W., Fink, D., Riedewald, M., Caruana, R., Ballard, G. and Hooker,

G. (2009) Data-intensive Science: A New Paradigm for Biodiversity Studies. BioScience

59(7): 613-620

Kitchin, R. (2014a) The real-time city? Big data and smart urbanism. GeoJournal 79(1): 1-

14.

Kitchin, R. (2014b) The Data Revolution: Big Data, Open Data, Data Infrastructures and

Their Consequences. Sage, London.

Kitchin, R. (2015) The opportunities, challenges and risks of big data for official statistics.

Statistical Journal of the International Association of Official Statistics

Kitchin, R. and Dodge, M. (2011) Code/Space: Software and Everyday Life. MIT Press,

Cambridge, MA.

Kitchin, R., Lauriault, T. and McArdle, G. (2015) Knowing and governing cities through

urban indicators, city benchmarking and real-time dashboards. Regional Studies, Regional

Science 2: 1-28.

MacKenzie, D. (2008) An Engine, Not a Camera. How Financial Models Shape Markets.

MIT Press, Cambridge, MA.

Miller, H.J. (2010) The data avalanche is here. Shouldn’t we be digging? Journal of

Regional Science 50(1): 181-201.

18

Mims, C. (2013) Coming Soon: The Cybercrime of Things. The Atlantic, 6th August,

http://www.theatlantic.com/technology/archive/2013/08/coming-soon-the-cybercrime-of-

things/278409/ (last accessed 7th August 2013)

Morozov, E. (2013) To save everything, click here: Technology, solutionism, and the urge to

fix problems that don’t exist. New York: Allen Lane.

Openshaw, S. (1984). The Modifiable Areal Unit Problem. Concepts and Techniques in

Modern Geography 38, Geo Books, Norwich.

Paganini, P. (2013). Israeli road control system hacked, causes traffic jam on Haifa highway.

The Hacker News, October 28th. http://thehackernews.com/2013/10/israeli-roadcontrol-

system-hacked.html. (Last accessed 13 Nov 2013).

Parsons, W. (2004) Not just steering but weaving: relevant knowledge and the craft of

building policy capacity and coherence. Australian Journal of Public Administration 63

(1),43–57.

Rameriz, E. (2013) The privacy challenges of big data: A view from the lifeguard’s chair.

Technology Policy Institute Aspen Forum, 19th August,

http//ftc.gov/speeches/ramirez/130819bigdataaspen.pdf (last accessed October 11th 2013)

Raley, R. (2013) Dataveillance and countervailance, in Gitelman, L. (ed) “Raw Data” is an

Oxymoron. MIT Press, Cambridge, pp 121-146.

Rameriz, E. (2013) The privacy challenges of big data: A view from the lifeguard’s chair.

Technology Policy Institute Aspen Forum, 19th August,

http//ftc.gov/speeches/ramirez/130819bigdataaspen.pdf (last accessed October 11th 2013)

Söderström, O., Paasche, T. and Klauser, F. (2014) Smart cities as corporate storytelling.

City 18(3): 307-320.

Stroud, M. (2014) The minority report: Chicago's new police computer predicts crimes, but is

it racist? The Verge, February 19th. http://www.theverge.com/2014/2/19/5419854/the-

minority-report-this-computer-predicts-crime-but-is-it-racist

Townsend, A. (2013) Smart Cities: Big Data, Civic Hackers, and the Quest for a New

Utopia. W.W. Norton & Co, New York.

Vanolo, A. (2014) Smartmentality: The Smart City as Disciplinary Strategy. Urban Studies

51(5) 883–898.

Wolfram, M. (2012). Deconstructing Smart Cities: An Intertextual Reading of Concepts and

Practices for Integrated Urban and and ICT Development. In Schrenk, M., Popovich,

V.V., Zeile, P. and Elisei, P. (Eds.), Re-Mixing the City: Towards sustainability and

resilience? REAL CORP. pp. 171–181.